- Restore the always-on blink/pulse (still motion-reduce:hidden for a11y).
- Drag the bubble off the left edge to hide it; a tiny left-edge tab brings it
back. (The ✕/panel/Settings Hide paths remain.)
- Replaced the cramped inline "Go Pro" header pill with a clean full-width
strip at the top of the panel body.
Sasquatch (the look regressed): I over-corrected last pass by shrinking the
bubble on the map and bolting an always-visible ✕ over the avatar on mobile.
Restored the full-size bubble + glow on every route (the map only loses the
pulse), moved Hide to a desktop-hover ✕ and an EyeOff "Hide assistant" button
in the panel header (Settings → Appearance still re-enables it).
Mints: the add-mint dialog only offered a hand-picked "Recommended" shortlist.
It now also lists the full searchable MintRadar directory (useMintDirectory)
below the recommendations.
Events: CreateEventPage could only link a skate spot. Added an optional
"Link to a Shop" picker (useSkateShops, search, location autofill) that writes
['a','30017:…'] + ['venue-name', …].
Addresses "the floating Sasquatch flashing is annoying":
- Pulse is now state-driven: the animate-ping halo shows only when there's an
unread tip (message), never perpetually; the unread dot matches.
- Honors prefers-reduced-motion (motion-reduce:hidden).
- Map-aware: Layout passes isFullBleed, so on the map the bubble runs compact
(smaller, no glow, no pulse) instead of competing with map controls.
- Hideable: a ✕ on the FAB hides it (useSasquatchHidden, localStorage
sasquatch-hidden, same-tab synced via an event), with a "Show Sasquatch
assistant" toggle in Settings → Appearance to bring it back.
- New src/lib/dateFnsI18n.ts re-exports date-fns with formatDistanceToNow /
formatDistanceToNowStrict / formatDistance / format defaulting to the active
locale, plus dateLocaleCode() for native toLocale{Date,Time}String calls.
- LanguageProvider keeps the locale in sync during render, so the first paint is
already in-language.
- Codemod: 11 date-fns imports now use the shim; ~39 native date/time calls pass
the app locale code instead of undefined/[]/hardcoded 'en-US'. Number
.toLocaleString() calls are deliberately untouched.
- First-visit language auto-detect (Spanish devices land in es; stored choice wins).
- Sasquatch AI replies in the app language (skate trick names kept English).
- i18n safety net: scripts/check-i18n.mjs wired into npm test — fails on any t()
key missing from en/es or any en/es key drift.
- ~1,680 new keys (2,297 total) across mapUi / spotUi / marketUi / labUi and
~65 files wired to t(): Map, spots (detail/create/edit/chat), marketplace,
trades, offers, escrow, shipping, The Lab, /become/* forms, coaches, skaters.
- Policy: only skate TRICK names stay English; user-generated content untouched.
Three phone issues from a screenshot review:
- Map: the floating toolbar sat on top of Leaflet's top-left zoom control.
index.css now offsets .leaflet-top.leaflet-left by 4.25rem under max-width
767px, scoped to the map page via a new class
(desktop toolbar is in flow, so no offset). Verified at 390x844.
- Sasquatch: the cramped red 8px 'Go Pro' chip is now a proper lime pill with
the bolt, matching the panel.
- Directions: the map was 38vh and buried under the mode row. It's now 50vh
(440px on larger screens; 60vh/62vh while following) and sits directly under
the header, with the mode/vehicle row and 'From' row moved BELOW it. Done with
flex order-* utilities rather than moving JSX; Start bar stays pinned.
1) Proximity intent: 'closest / near me / nearby' + spots is handled ON DEVICE —
the fix is read locally (fresh, else cached) purely to sort distances, and
nothing is transmitted (not to relays, not to the AI). Answers the nearest 3
with distances and says location never left the device.
2) System-prompt guardrail: the AI CANNOT see the user's location — for
closest/near-me it now says so and points at the map's Near Me (or asks for a
city), instead of guessing.
3) Intent hijack fix: a question mentioning both a crew and a spot was routed to
find_crews ('crews' was checked before 'spots'). Spots now win when both
appear.
- Header: name + Pro badge on line 1; context title + (owner-only) AI balance
on line 2. The live/CVM chip and AI balance no longer fight for the same row
and wrap.
- Footer: the live-status chip now sits with 'Live data from Nostr relays' and
'clear' instead of the header.
- Input: taller (h-10) with text-sm on mobile; send button matches.
- Message bubbles: text-xs and roomier padding/radius.
- #3 usePersonalContext decrypts the Stash ON DEMAND (unlockStash, called by
getMyStash) instead of never — 'what's in my stash?' is now correct, while the
globally-mounted guide still never decrypts on mount. stashRef avoids the
stale-closure bug.
- #4 new self-contained action draft wtb_gear (NIP-99 listing_type=wtb) — a
'wanted' gear ad. SYSTEM_PROMPT + tool handler + publish branch.
- #5 panel header shows 'Pro · Nd' days left (title = exact expiry).
- #6 tier picker computes live USD via useExchangeRate; removed the hardcoded
~$60k/BTC usd field from PRO_TIERS.
- #7 copy: 'skate status' not 'spot status'; trip planning stated as free.
#1 Extend now actually extends. useSasquatchPro collected receipts and took the
max of (receipt.created_at + tierDays), so buying again before expiry
under-delivered. It now stacks every valid proof with renewal semantics
(expiry = max(expiry, proofTime) + days), client and server alike.
#2 Owner-grant fallback. Pro can now also be granted by an owner-authored NIP-32
kind-1985 label (org.skatespots.pro, p = user, expiry) — so Pro is not solely
dependent on the Lightning provider publishing zap receipts. New owner-only
GrantProCard (Settings -> Advanced) publishes it to getWriteRelays(); the hook
reads via querySkateContent.
Protocol: the x-sasquatch-pro payload is now {receipts,grants}; BOTH ai-proxy and
the Netlify function accept the new payload and a bare receipt (old APKs). Proxy
was deployed first and verified healthy before this commit.
Owner-only chip (next to the Pro/CVM chips) showing the upstream sats balance,
amber when low. useSasquatchAiStatus gains an 'enabled' option; the panel only
fetches while it is open, so the NIP-98 signature never fires at app startup.
- netlify/functions/ai.mjs: optional Pro enforcement (SASQUATCH_REQUIRE_PRO=1).
Requires a NIP-98 kind-27235 auth header (x-sasquatch-auth) and, unless the
caller is a platform owner, an active Pro pass (x-sasquatch-pro = a
provider-authored zap receipt) verified OFFLINE — signature, user binding,
platform tag, and paid-amount → tier. Off by default.
- src/lib/proAuth.ts: signs the NIP-98 once per session (no Amber popup per
message) and attaches the Pro receipt.
- sasquatchBrain: threads auth headers through the proxy (streaming + non).
- useSasquatchPro: exposes the winning receipt event.
- Verified: 200 with the gate off, 401 with it on (env reset afterwards).
- useSasquatchPro: derive Pro expiry from what was ACTUALLY paid (largest tier
the bolt11 covers, run from the receipt time), instead of trusting the
user-signed `pro-expiry` tag — a 7-day payment could previously self-assert a
year. Provider-authored receipt + amount checks unchanged.
- SasquatchGuide: AI answers are now Pro-only. Free users get the static FAQ,
glossary, trip planner and local relay lookups (all no-cost), plus a nudge to
Pro when a question would otherwise need AI.
- AGENTS.md updated.
- sasquatchBrain: new streamSasquatchAI — OpenAI-style SSE via OpenRouter (if
the key is set) then Pollinations; falls back to the non-streaming caller only
when nothing at all was delivered, so blocked-stream environments still answer.
- SasquatchGuide: stream the first pass into a live bubble (raw JSON tool calls
are suppressed so partial JSON never shows; prose-then-tool output is dropped),
finalise the bubble in place, show a blinking cursor, and hide the generic
'thinking...' bubble while a reply streams.
- Drop Pollinations 'mistral' (now a 404) leaving the live 'openai' slug.
- sasquatchBrain: abort each provider request after 15s and give up the whole
multi-provider attempt after 35s, so a hung OpenRouter/Pollinations call can no
longer wedge the chat (it now falls through to the static/local fallbacks).
- SasquatchGuide: send the system prompt as a real system message (both the
first pass and the tool-synthesis pass) instead of smuggling it into a user
turn, and mark tool output as untrusted — blunts prompt injection via event
content.
- useCVMClient: signature-verify (verifyEvent) every relay result, matching the
standalone server and the release-notes claim.
- useCVMServerHealth: stale threshold 1h -> 45m to match the server heartbeat.
Audit found the Pro paywall promised three things with no implementation:
- "Priority AI model, higher limits & longer memory" — Pro and free hit the
exact same model/limits/history (only difference is the personal-tools prompt)
- "Proactive nudges (weather + spot tips)" — no nudge/scheduler exists anywhere
- "Pro-only Sasquatch skins" — single hardcoded avatar; no skin system
- "Supporter badge" — gating exists but no badge is rendered
Rewrote all Pro copy to promise ONLY what actually ships (personal assistant
over the user's own Stash/crews/spots/listings, confirm-before-publish action
drafts, personal-context answers, and directly funding an ad-free app):
- SasquatchProDialog PERKS list (dropped Sparkles import now unused)
- usePurchasePro receipt DM
- 6 STATIC_FAQ Pro answers in SasquatchGuide
Pricing, payment target (skatespots@minibits.cash) and npub-bound zap-receipt
verification were already accurate and left unchanged. Build clean, no console
errors.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
The chat's data layer (useCVMClient) queries Nostr relays directly from the
browser — the CVM/ContextVM server round-trip was removed earlier for speed.
But the UI and system prompt still claimed data came "from the CVM server",
which is no longer true. Aligned every user-facing claim with reality:
- System prompt: reframed "CVM Tools / MUST route through CVM" as "Live data
tools that query the Nostr relays directly". Tool-calling behaviour unchanged.
- Streaming status: "Querying CVM for X" → "Searching the relays for X".
- Tool-failure message: no longer blames a non-existent "CVM server"; now
correctly frames it as relays being slow, with a retry nudge.
- Empty results: stop asserting "not added yet" as fact — an empty result can
also mean slow relays or a misspelled place. Now offers retry + add-a-spot,
and names the place searched. Applied to the local formatter, the system
prompt rules, and the second-pass AI instruction.
- Chat footer: "Powered by ContextVM & Nostr Relays Data" → "Live data from
Nostr relays" (links the primary community relay).
- About page credit: "AI powered by ContextVM" → "Trust & crew chat via
ContextVM" (its real role: Relatr trust scores + encrypted crew chat).
- Owner CVM health panel: clarified it probes the standalone MCP server and
that the in-app chat does not depend on it.
- Updated stale internal comments in SasquatchGuide + CreateSpotPage.
ContextVM references that ARE accurate (Relatr trust scores, crew-chat Cordn
coordinator) were left intact. Build clean, no console errors.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
The local fast-path was stricter than the AI it replaced: a typo like "bangok"
won't substring-match "Bangkok", so the fast-path returned a false "No spots
listed there yet" even though Bangkok has spots.
Two-layer fix:
1. Safety net (most important): if a finder fast-path comes back EMPTY, fall
through to the AI path instead of rendering "nothing found". The AI normalizes
spelling and can still find results. Stats/weather always render immediately.
Net effect: fast when the local match works, correct-always via AI when it
doesn't — never a false empty.
2. Typo tolerance: findSpots city matching now uses placeMatches() — substring
OR edit-distance-1 against candidate word tokens (only for queries ≥4 chars,
so short words can't over-match). So "bangok" matches "Bangkok" on the fast
path too. Freeform content stays a strict substring check to avoid false hits.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Root cause of the ~35s "spots in bangkok" response: every live-data question
first waited on an AI round-trip just to produce a tool call, and the keyless
backup model (Pollinations) can take tens of seconds. The relay query and local
formatting were already fast — the AI was the bottleneck.
Fix: before touching the AI, try parseLocalIntent(). For clear finder questions
(spots/events/coaches/crews/shops/tours/gear + community stats + weather with a
parseable location) we now go STRAIGHT to the relays (capped at 4.5s) and render
locally — zero AI calls. Expected ~35s → ~2–4s on the most common queries.
Deliberately still routed through the AI:
- trip planning (needs the rich itinerary card)
- personal "my …" Pro queries (get_my_* tools)
- anything without a confidently-parseable intent/location
If the fast-path relay query errors, it falls through to the existing AI loop as
a backup, so nothing regresses.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Two bugs from "spots in bangkok" (said "Found 20" but showed 6, with repeats):
1. Misleading count: formatLocalResults hard-capped the visible list to 6 while
the headline reported the full count. Now shows up to 10 and appends a
truthful "…and N more — see them all on the map" line, so the number and the
list always agree.
2. Duplicate spots: coordinate dedupe (kind:pubkey:d) can't catch the same
real-world spot re-posted under a different identifier or by another account
(e.g. "Yellow Park – Sukhumvit 50" appearing twice). Added
deduplicateByIdentity() — a second-pass dedupe on a normalized
name + geohash/location signature, keeping the freshest event — applied in
findSpots so the count and list are clean.
Both fixes flow through the AI tool-call path and the offline fallback path.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Three low-risk speedups focused on the slowest path — questions that hit live
relay data:
1. Relay timeout 10s → 4.5s (6s for the fan-out community stats). The useful
relays answer in ~2s; the old 10s just let one slow/dead relay stall the
whole reply.
2. Results-first: for finder tools (spots/events/coaches/gear/crews/shops/
tours/stats/weather) format the CVM result locally and SKIP the second
"prettify" AI call. Removes a full model round-trip (~2–4s) — users want the
list, not extra prose. New toolCallToIntent() maps a tool call to the existing
formatLocalResults() path. Personal/Pro tools (get_my_*) still use AI synthesis.
3. Faster AI fallback: cap OpenRouter attempts to the top 3 free models before
the keyless backup, and shorten 429 backoff (≤800ms vs ≤2500ms). Avoids the
worst-case 15–20s grind through 8 models on a busy free tier.
Conversational/knowledge answers (single AI call) keep Sasquatch's personality
and are untouched.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
- Panel now anchors md:bottom-16 md:left-4 (was bottom-right) so it opens on
the left as before, and sits ABOVE the desktop footer so it never covers the
"Made with..." row or footer links.
- Tightened max-height to clear the top nav too (never slides under the menu).
- Bubble stays bottom-right on desktop for discoverability (clear of footer).
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
- findCoaches now covers EVERY Lab craft (coach, filmer, editor, photographer,
builder, artist, graffiti, tattoo, trainer, musician, judge) via role param;
reads skills from `t` tags and strips self-tags.
- New findShops (kind 30017, t=skateshop) and findTours (kind 30823) tools.
- getCommunityStats counts all crafts individually + lab_total + shops + tours,
and filters spots to t=skatespot (drops foreign kind-30818 events).
- Wired new tools + expanded role vocabulary into the Sasquatch brain: system
prompt, intent parser, keyword detector, and both dispatch/format paths.
- Privacy preserved: the public CVM data layer never queries the encrypted
Stash (kind 30001) — verified no 30001/decrypt references. Personal 'my
stash' stays login+Pro gated and decrypts locally with the user's own key.
- Normalized useCVMClient.ts to clean UTF-8 (comment glyphs only; no code change)
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
- Move bubble & panel to bottom-RIGHT on desktop (was bottom-left, colliding
with / hiding behind the desktop footer). Mobile stays bottom-left.
- Raise bubble/panel z-index to z-60 so it always sits above the footer.
- Re-clamp a stale saved drag position on mount: a bubble dragged off-screen
on one device no longer disappears on another.
- Stop showing the long welcome greeting + action button on every open once a
conversation has started (chatMessages > 0), so it no longer "says the same
thing all over again".
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
- NofferField: link → https://clinkme.dev/specs.html (instead of just /)
- EditProfileForm: link → https://clinkme.dev/specs.html (instead of just /)
- SasquatchGuide: added [specs] and [apps] inline links next to the clink description, pointing to the new pages
The app already referenced clinkme.dev homepage; these add the two new deeper pages for discoverability.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
"Start a Crew" was ambiguous (team? chat? listing?). Renamed to the clearer
"Create Crew" across the Crews page button/empty state, Create Crew page
title/header/login gate, and Sasquatch FAQ. Added a subtitle distinguishing a
Crew (skate WITH) from The Lab (hire).
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
1. Pro purchase confirmation: on a paid pass, send the buyer an encrypted
NIP-17 receipt to their own inbox (tier, expiry, unlocked perks). Best-effort
so it never blocks the purchase. Success toast now points to the DM.
2. Frictionless Pro answers: add local static-FAQ entries for "sasquatch pro",
"pro perks/version", pricing, "go pro", "upgrade" — so those questions are
answered instantly and never dead-end when the free AI model is rate-limited
(the exact failure seen when asking "what perks has the pro version?").
3. FAQ matching now prefers the longest matching key so specific phrases win
over broad ones ("sasquatch pro" > "pro").
4. Refresh the free-model fallback list (add deepseek-v3, gemini-2.0-flash-exp,
qwen-2.5-72b) for better free-tier AI availability.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
- Trip Planner: `plan_trip` tool orchestrates spots+events+coaches+crews+weather
into a weather-aware, day-by-day itinerary card (TripItineraryCard). Actions:
save all spots to a NIP-51 Collection, book/zap a coach via their profile,
and deep-link out to flights/stays (assist, never fake-book, no KYC).
- Free-tier resilience: local intent parser answers "spots/events/coaches/
crews/gear/weather in <city>" and full trip plans directly via CVM even when
both AI providers are rate-limited — no more dead-end fallback.
- New shared joke book (sasquatchJokes.ts): skate humor + gentle roasts of the
KYC-your-audience Bitcoin-influencer crowd. Idle joke on the empty chat state,
ethos jokes in fallbacks. Dedupes the old inline joke arrays.
- UX polish: taller chat scroll area, trip-planner suggested questions, FAQ
entries for the planner.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Sasquatch can now DRAFT actions from natural language and publish them
only after the user reviews & confirms — it never publishes on its own.
- useSasquatchActions: publishes confirmed drafts via the app's normal
publish path. Supports post_status (NIP-38 kind 30315, auto-expiring)
and list_gear (NIP-99 kind 30402, skateboarding-tagged), reusing the
existing event schemas.
- SasquatchActionCard: inline "review & confirm" card with editable key
fields (status text / listing title, price, condition), Publish +
Cancel, and done/cancelled states. Nothing is signed until Publish.
- SasquatchGuide: new draft_action tool (Pro-gated). The AI emits a
draft; a card renders in-chat; confirmDraft signs & publishes. Non-Pro
users get an upgrade nudge. Draft lifecycle tracked per message.
- Prompt teaches the model to use draft_action for "post/sell/list/set
my status" requests and to fill gear fields from the stash first.
Examples now working (Pro): "post that I'm skating Southbank for 3 hours"
· "list my Spitfire wheels for 8000 sats".
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Adds the Sasquatch Pro supporter pass with strong, forge-resistant
verification bound to the paying user's npub — no backend required.
Verification (useSasquatchPro):
- STRONG: NIP-57 zap receipt (kind 9735). Purchase signs a zap request
tagged pro-tier/pro-expiry; the receipt embeds it, cryptographically
binding payment → identity (unforgeable by the user).
- FALLBACK: self-signed kind-32671 receipt so paying users are never
locked out if the provider emits no zap receipt.
- Platform owners always Pro. Access = max unexpired expiry.
Purchase (usePurchasePro):
- Zap to skatespots@minibits.cash (falls back to plain invoice), pay via
NWC → WebLN → manual QR. Publishes the fallback receipt on success.
Pricing (sasquatchPro.ts): 4 tiers — 7d/2k, 1mo/5k, 3mo/12k (popular),
1yr/36k sats.
UI (SasquatchProDialog): upgrade sheet with tiers, perk list, QR
fallback, "already Pro" state, extend flow.
Gating (SasquatchGuide):
- Personal tools (get_my_*) now require login + Pro; prompt only exposes
them to Pro users; tool switch guards with a pro_required nudge.
- ⚡ Pro badge when active, ⚡ Go Pro button otherwise; local upgrade
nudge for "my/mine" questions from non-Pro users.
Docs: kind 32671 + verification model documented in NIP.md.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Sasquatch can now answer questions about the logged-in user's OWN data
("what's in my stash?", "my crews", "events near my crew", "my spots",
"what am I selling?").
- New usePersonalContext hook: client-side getters for the user's Stash
(decrypted locally), crews (founder/member), marketplace listings, and
created spots. Reads own data only — nothing personal leaves the device.
- Registered 4 personal tools (get_my_stash / get_my_crews /
get_my_listings / get_my_spots) in the Sasquatch tool switch.
- Personal-tools prompt block is appended to the system prompt ONLY when
a user is logged in, so logged-out UX is unchanged.
- needsLiveData now routes "my/mine" questions to the AI tool path.
Phase 1 is intentionally ungated so we can validate the experience
before wrapping the Pro paywall (Phase 3).
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Fixes the "brain's catching traffic" dead-end for trick/terminology
questions and reduces reliance on a single flaky free API.
- Local glossary answerer (answerFromGlossary): trick/how-to/terminology
questions ("what is a feeble grind", "how do I do a kickflip", "grosso
grind") are now answered instantly from the 150+ entry glossary with
ZERO API calls — so they never hit a rate limit.
- Wired into SasquatchGuide as (a) a fast path before the AI and (b) a
last-resort fallback when the AI is unreachable.
- sasquatchBrain now has multi-provider fallback: OpenRouter free models
(with 429 backoff) → keyless Pollinations backup. Only throws if both
fail, letting the caller fall back to local knowledge.
- Replaced the unhelpful fallback message with useful local links +
a nudge to ask about a specific trick.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
Glossary:
- New /glossary page: 80+ tricks across 11 categories (fundamentals,
flips, shoves/spins, grinds, slides, grabs, airs, lip, stalls,
handplants, boneless, freestyle) with alternate names, difficulty
badges, live search, and category filters.
- Added to desktop "More" dropdown and mobile burger menu.
Sasquatch AI:
- Fixed/expanded free-model fallback list with valid OpenRouter slugs
(dropped the non-existent gemma-4 slug).
- Added exponential backoff + one retry on 429 rate limits, respecting
Retry-After headers.
- SasquatchGuide now reuses the shared sasquatchBrain caller instead of
a duplicated implementation.
- Friendlier fallback message when the AI is saturated.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
The "1st tap does nothing, 2nd/3rd works" bug came from relying solely on
pointer-capture tap detection, which is flaky on the first interaction after
mount in mobile WebViews — and the earlier e.preventDefault() on pointerup was
suppressing the click we could otherwise use.
Fix: add a plain onClick as a reliable open path (the most dependable "tapped"
signal in a WebView), extract an idempotent openPanel() used by both pointerup
and click, and drop the preventDefault so the click isn't swallowed. The
justOpenedRef lock still blocks the ghost-click that would re-close it. Double
firing is harmless since openPanel just sets visible=true.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
On touch devices, the bubble's pointerup opened the panel and unmounted the
bubble; the synthesized "ghost click" from the same tap then landed on the
panel's minimize/close button (which renders in roughly the same bottom-left
spot), instantly re-closing it.
Fix: preventDefault/stopPropagation on pointerup to suppress the ghost click,
guard releasePointerCapture against unmount throws, and add a 400ms
"just opened" lock so stray close/minimize events from the same tap are
ignored.
Co-authored-by: shakespeare.diy <assistant@shakespeare.diy>
